Should you publish or patent first when applying for an NIH SBIR? File your provisional patent first, then publish. Publishing before you file destroys your patent rights in Europe, China, and most countries outside the US, which has a one-year grace period. A preprint posted after your provisional closes the grant's publications gap without sacrificing your IP.
That is the whole answer. The rest of this piece explains why it works, what happens if you get the order wrong, and exactly how to sequence the four steps.
Here is the trap. When a biotech founder runs a grant readiness assessment, "no publications" often comes back as the single biggest competitiveness gap for NIH. The instinct is to rush something into print. But the fastest way to publish is also the fastest way to lose your patent in most of the world. The two goals collide on timing, and the collision is avoidable.
Why "no publications" costs you 15 points at NIH
In Cada's grant roadmap scoring, a missing publication record is the largest single-variable penalty we apply. Most agencies dock 5 points for no peer-reviewed publications. NIH docks 15.
That is not a rounding error. Programs are scored on a 0-100 internal fit scale, and anything below 50 after modifiers gets cut from the roadmap entirely. A 15-point hit can move a company that was otherwise competitive (say, a 58) straight into the excluded pile (43). No other single gap does that.
Why does NIH weight publications so much harder than everyone else? Because NIH study sections are panels of working scientists. They read your publication record as a proxy for whether your science is real and whether your team can execute. A peer-reviewed paper, or even a preprint, signals that someone outside your company has seen the work and it survived contact with reality.
Compare that to where NIH does not penalize you: NIH applies no penalty for missing patents. The patent penalty (-5) shows up at NSF and DoD programs, which care about commercialization and defensibility. NIH cares about the science. That asymmetry is the root of the whole problem. The thing that helps you most at NIH (publishing) is the thing that can hurt you most on IP.
Why publishing first can destroy your patent
Here is the part founders miss, and it is expensive.
The United States gives inventors a one-year grace period. Under 35 U.S.C. 102(b)(1), your own public disclosure does not count as prior art against you, as long as you file your patent application within 12 months of that disclosure. So in the US, you can publish and still patent, if you move within a year.
Most of the rest of the world does not work that way. Europe, China, Japan, and the majority of patent systems follow an "absolute novelty" standard. There is no grace period. One public disclosure before your filing date, and your right to patent that invention in those countries is gone. Permanently. There is no fixing it later.
For a biotech building toward a global market, foreign patent rights are not optional. Losing Europe and China on your core invention can quietly erase most of the company's enterprise value, and you would not feel it until a Series B diligence call years later.
What counts as a public disclosure for patents?
A public disclosure is any non-confidential sharing of the invention that lets a person skilled in the field understand it. That includes a journal article, a conference talk, a poster session, a published abstract, a thesis on a library shelf, a product demo, and a detailed pitch deck shared without an NDA.
It also includes a preprint. This is where founders get burned.
A bioRxiv or medRxiv preprint feels informal. It is not peer-reviewed, it went up in a few days, it has no journal masthead. Founders treat it like a blog post. Patent law does not. A preprint is a public, permanent disclosure with a DOI and a timestamp. It is full prior art, worldwide, from the day it posts. Posting a preprint before you file a provisional is exactly as damaging to your foreign rights as publishing in Nature.
The sequence that protects both
The fix is not choosing patent or publication. It is ordering them. Four steps:
-
File a provisional patent application first. This establishes your priority date and starts a 12-month clock. The USPTO filing fee for a small or micro entity is a few hundred dollars or less, and a provisional drafted by a patent attorney typically runs $2,000 to $5,000. (Fees change, so confirm current numbers.)
-
Post a preprint to bioRxiv or medRxiv. It is public and citable within days, not the 6 to 18 months peer review can take. The preprint closes your publications gap roughly 10x faster than waiting for a journal.
-
Cite the preprint in your NIH application. NIH has explicitly allowed this since 2017. Under the interim research products policy (NOT-OD-17-050), you can cite preprints anywhere you cite other research products, including your biosketch and your preliminary data narrative. Study section sees a citable, dated, public research product with your name on it.
-
File your non-provisional or PCT application within 12 months of the provisional. This locks in your foreign rights. Because it claims priority back to your provisional, your effective filing date is before the preprint went public.
That last point is the whole trick. Your foreign patent filings claim a priority date that predates any public disclosure. The preprint cannot be prior art against your own application, because your application is dated earlier. You close the grant gap and keep your worldwide IP.
One honest caveat. A provisional only protects what it actually describes. If your preprint reveals more than the provisional did (new data, a new method, an additional embodiment), that extra material is not covered by your priority date and can become prior art against the uncovered parts abroad. So the rule is: draft the provisional to cover everything the preprint will say, then post the preprint. Not the other way around.
What if you have already published?
Damage control depends on timing.
If you disclosed within the last 12 months: your US rights are probably still alive. File a patent application now, inside the grace period, before the clock runs out. Your foreign rights in absolute-novelty countries are most likely already gone for whatever you disclosed, but the US market alone is often worth filing for.
If you disclosed more than 12 months ago: that specific invention is likely unpatentable, in the US and abroad. That is the bad news. The better news is that improvements, new methods, next-generation versions, and unpublished aspects of the work may still be patentable. File on those.
This is the situation where a patent attorney earns the fee several times over. The cost of guessing wrong here is the patent itself, so do not run this calculation alone.
What this actually costs you
Run the math, because it makes the decision obvious.
A provisional patent costs a few thousand dollars at most. A preprint is free. Sequencing them correctly costs you maybe two to three weeks of lead time before your grant deadline.
Getting it wrong costs you a Phase I award you were otherwise competitive for, or your patent rights across most of the global market, or both. NIH SBIR Phase I awards run around $314K. Foreign patent rights on a core biotech invention can be worth orders of magnitude more. A few thousand dollars of provisional filing is cheap insurance against losing either.
There is one case where you should not publish at all: a pure trade-secret play. If your edge is a manufacturing process, a proprietary dataset, or a method that is hard to reverse-engineer, a preprint hands your competitors a roadmap. Sometimes the right move is to eat the 15-point penalty, target agencies that do not weight publications as heavily, or build competitiveness another way. That is a real tradeoff, and the honest answer is that it depends on what your moat actually is.
Closing the gap the right way
If your roadmap flagged "no publications," you do not have to choose between a competitive grant and your patent position. You have to sequence them: provisional first, preprint second, grant third, full patent filing inside the year.
The hard part is not the concept. It is coordinating the patent attorney, the preprint, and the application timeline against a fixed NIH deadline, while drafting a provisional broad enough to cover what you are about to publish. That coordination is your SBIR publication strategy, and it lives or dies on sequencing.
If you want the broader picture of how different agencies treat IP and publications, the same logic extends past NIH: NSF and DoD weight patents, foundations and prizes weight neither. (See our agency IP comparison matrix for the full breakdown.)
That coordination is what we do. Cada builds NIH SBIR applications with a patent-publication timeline mapped to your filing strategy, so closing the competitiveness gap never costs you your IP. If you want a straight answer on the right sequence for your specific situation, we do a free publication strategy consultation. No pitch, just the sequence.
One thing to be clear about: Cada is not a law firm, and nothing here is legal advice. We coordinate the grant strategy around your IP; your patent attorney makes the filings. Get both in the same room early, and the timing trap stops being a trap.